SGLang

SGLang is a model-serving engine with an OpenAI-compatible service surface and its own runtime-native graph/DSL features. LychD binds the declared service through the Soulstone adapter; it does not require the caller to depend on sgl.gen or any other SGLang-specific language.

The current adapter uses operator-owned exec passthrough: the Rune supplies the complete SGLang launch command, while typed fields retain container and endpoint intent. Its registered OpenAICompatibleRuntimeAdapter binds that declared service; it does not synthesize an SGLang program from a generic endpoint. Keep the command's listening port and served model aligned with the declared endpoint and model identity.

Pin framework flags with that command and include them in the exact engine profile and runtime receipt. Neither a healthy URL nor a generic compatibility label establishes every declared model or API operation.

The exact image, model, quantization, GPU and driver, launch arguments, readiness, inference, and shutdown still require an operator receipt. The current delivery boundary is recorded under SGLang integration.

See Soulstone Disciplines, SGLang, and the OpenAI-compatible connector boundary.